HF: 0.5W Linear push pull PA

@10/7/2015

i did build some PA, almost all of them did not meet expectation, that is just experiment to find way to understanding a PA. from a bug transmitter(several years ago)  to something called transmitter,  these testing and building gain some experience, let me have a instinct of how the PA works in that way.

when built  AM band transmiter (Techlib)  i totally don't understanding what happens there, and there is no instrument to do any  kind of check. that AM transmitter built 2 years ago, when i review that today, i recall some details, that is it's never seems to properly match to that antenna anyway.

same year with that AM band transmiter, i tried to build a CB band transmitter, 27Mhz: AM RX/TX  , at that time the expectation is a transmitter can emit hundreds mW to air, details recording in another little article Exploring PA , that AM modulator is weak without a audio preamplifier, and output obvious could not that powerful with one lipo battery except 8.4V to 9V. that little transmitter does could given about 200mW output, but a class C PA for AM is totally wrong.  

except that, there is lots of error in design, like, lack understanding the match networking, driven stage to final PA impedance did not properly matched. the driven stage to final PA impedance match is critical to achieve better result.  class C PA need a little bit strong driven signal, most important thing is driven stage could given a 2Vpp signal to 50ohm load does not mean it could driven a class C PA, cause BJT's input impedance roughly just about few ohm.

checking the data in Exploring PA, there is lots of suspect data there, i.e, driven stage deliver to a 50ohm load get lower voltage peak  to peak value than  it driven the final PA, this does not make sense, cause final PA's input impedance is much more lower than 50ohm. the reason is unknown now, mostly probably because testing method is totally wrong. but that project given me a clue about the class C PA, that's need a good driven stage properly impedance match and there is a minimal driven signal level required, and Class C  is not suitable for AM transmitter. 

last year, 2014, i start another experimental project, trying to design a TX chain for a FM CB band transmitter, PA: TX chain smoking Design. there, i did achieve something, and data present there is reasonable, driven stage deliver to 50ohm load is 2.5Vpp, when driven to final PA, reduce to 1.2Vpp. that driven stage use a class A design and that transformer is not very suitable because it's had  low XL and driven signal level is still not enough, but, better than that AM transmitter, get 360mW output finally.

these type of core, almost can not working in HF i guess, i pick up them in the surplus market, it's very very cheap, maybe could only work at MF, and even low frequency.  20T on those core, you will get a 1mH inductor. 

the truth is , this type of core works very well at HF, i get 600mW this time.

* the transistor is a little bit hot, i thought it could because there is no glue to fix the 2N2222 to copper ground for good heat dissipation

* output signal a little bit distortion, might be over driven

compare to PA which i design and made， there is several significant difference:

 * the MOST big difference is the match networking,  T1 is 11T:1T for a single PUSH PULL transistor  that is 121:1 impedance transformer, present to T1 collector pretty high impedance (roughly  1.2k impedance  11*11*10R, but need considering the T1's Q)

*  T1 use a 47R degradation resistor to achieve almost a 50 Ohm  input impedance.

the later 2 board i built, it running into problem.  the PA self oscillation, even with extra bypass and DC isolation. first time i think it might because over driven, i  then try to reduce driven by some attenuation, but, it did not help.

when i check the circuits, found if i touch the T1's Collector, the oscillation will stopped, this remind me, there is couple from T1's collector to base, make it unstable when driven harder.  the unwanted signal coupled  there, because i soldering the T1 with long lead, and collector is close to it's base. when i re-position  T1's collector and base,  it's stable. driven it to 400mW is fine, waveform is pretty everywhere.

i use a IF can as L1, a primary:secondary about 3.  T1 replaced with a 2N2222, Y1 change to a 9Mhz crystal, C3 exploit the internal capacitor inside the IF Can. to enable this oscillator working on 12V, change R2 to 13k.  T1's secondary connect to PA's input via a  0.1uF capacitor.  C2 is omitted.

i trying to cure this problem, first i think this because the wire from this oscillator to the PA input is too long, it is, about 5 cm. another reason might be the IF can pickup the strong signal. ground the IF Can but it does not help any a little bit. 

i found when i screw the top hat into the IF CAN,  the output could be stable but oscillator nearly stop oscillating. there is too much un certain issue in this oscillator. The Coil, the bias, even though this oscillator could output about 3 dBm to a 50R load, i don't think driving the PA directly from this oscillator is good. 

use a oscilloscope probe testing the T1's base waveform, i got a peak to peak voltage. RF output connect to a  Power Meter via a step attenuation. i try convert the peak to peak voltage to a input dbm,  then get the power Gain.  But this is the wrong method, cause input impedance is not exactly 50ohm at T1 base. 

Transducer Power Gain  is the most easy to get, connect my driven stage to the Power Meter, get 3 dBm, switch in the PA, get 27.5 dBm, so my amplifier GT = 24.5dB, not that bad. 

oscilloscope result:

BC547's base:  1.5Vpp

BC547's collector: 10Vpp (output power reduce 6dBm while hook the X10 probe to collector, so this is meaningless to calculate anything !!)

T1 secondary voltage: 2Vpp,  this is useful to calculate something if you want know push pull stage voltage gain, 12/2 about, 6 times voltage gain, about 15dB.